The platform design paradigm from the early days of Hadoop has been to co-locate compute and storage on the same server, which requires expanding both in tandem as your Hadoop cluster grows. This is an acceptable approach for workloads that need to expand compute and storage simultaneously. But, as Hadoop gained mainstream adoption, enterprises started finding workloads where the need for storage outpaced the need for compute. And now, after a decade of big data, enterprises are finding that historical data sets, though accessed less frequently, still need to be easily accessible. This has brought forth … READ MORE

Source: Dell Blog
—